Disruption of Gene pqqA or pqqB Reduces Plant Growth Promotion Activity and Biocontrol of Crown Gall Disease by Rahnella aquatilis HX2

Rahnella aquatilis strain HX2 has the ability to promote maize growth and suppress sunflower crown gall disease caused by Agrobacterium vitis, A. tumefaciens, and A. rhizogenes. Pyrroloquinoline quinone (PQQ), a cofactor of aldose and alcohol dehydrogenases, is required for the synthesis of an antibacterial substance, gluconic acid, by HX2. Mutants of HX2 unable to produce PQQ were obtained by in-frame deletion of either the pqqA or pqqB gene. In this study, we report the independent functions of pqqA and pqqB genes in relation to PQQ synthesis. Interestingly, both the pqqA and pqqB mutants of R. aquatilis eliminated the ability of strain HX2 to produce antibacterial substance, which in turn, reduced the effectiveness of the strain for biological control of sunflower crown gall disease. The mutation also resulted in decreased mineral phosphate solubilization by HX2, which reduced the efficacy of this strain as a biological fertilizer. These functions were restored by complementation with the wild-type pqq gene cluster. Additionally, the phenotypes of HX2 derivatives, including colony morphology, growth dynamic, and pH change of culture medium were impacted to different extents. Our findings suggested that pqqA and pqqB genes individually play important functions in PQQ biosynthesis and are required for antibacterial activity and phosphorous solubilization. These traits are essential for R. aquatilis efficacy as a biological control and plant growth promoting strain. This study enhances our fundamental understanding of the biosynthesis of an environmentally significant cofactor produced by a promising biocontrol and biological fertilizer strain.


Introduction
The gram-negative bacterium Rahnella aquatilis is widely ubiquitous, thriving in soil, water, marshes, and on food, seeds and plant roots. Strains of R. aquatilis fix nitrogen in the rhizosphere, solubilize mineral phosphate, and have biocontrol capabilities [1,2,3]. Specifically, R. aquatilis HX2 has been shown to suppress sunflower crown gall disease caused by Agrobacterium vitis, A. tumefaciens, and A. rhizogenes [4]. Biocontrol activity by R. aquatilis is nonspecific. This species has demonstrated suppression of diseases caused by Xanthomonas campestris, X. axonopodis, Penicillium expansum, Botrytis cinerea, and Erwinia amylovora [5,6,7]. Competitive colonization of varied environments and the ability of HX2 to establish several beneficial interactions with plants make it an ideal candidate for a soil inoculant. A mechanistic understanding of plant disease suppression and mineral phosphate solubilization by a R. aquatilis was provided in previous research, which linked these activities to the glucose dehydrogense cofactor, pyrroloquinoline quinone (PQQ) [8,9].
In gram-negative bacteria, PQQ mainly functions as a non-covalently bound, redox cofactor of several membrane-associated sugar and alcohol dehydrogenases, including methanol dehydrogenase, ethanol dehydrogenase, and glucose dehydrogenase (GDH) [10]. Previous reports indicate that the GDH-PQQ holoenzyme is involved in the production of an antimicrobial substance by several genera of bacteria including Rahnella and notably, strain HX2 [9,11,12]. Organic acids (OA) such as gluconic acid (GA) are considered to be a main factor responsible for dissolution of insoluble phosphate through organometallic complex formation or through metal chelation processes [13,14]. Bacterial mineral phosphate solubilization (MPS) activity most commonly occurs when bacteria produce and release OAs [14,15,16]. Furthermore, PQQ is a plant growth promotion factor, which in addition to GA production, has also been related to its antioxidant properties as well as unknown mechanisms [17].
Bacterial genes involved in PQQ biosynthesis have been identified in numerous species isolated from varying environments and are clustered in pqqABCDEF operons [9,18,19,20]. The pqqA gene encodes a small peptide that contains tyrosine and glutamate and serves as the precursor and rate-determining step for PQQ biosynthesis [21]. This molecule remains attached to a precursor peptide and is cleaved off at a later step by other enzymes of the biosynthesis pathway. Often, pqqB is not directly required for PQQ biosynthesis. Its suggested role in K. pneumoniae is a carrier that facilitates the secretion of PQQ across the plasmamembrane into the periplasm [22]. Little information is available on PQQ biosynthesis in the Rahnella genus. Kim et al. [8] mobilized a cosmid library of R. aquatilis into Escherichia coli HB101 to isolate and clone the genes that confer the MPS trait from R. aquatilis. Consequently, it was revealed that the MPS locus of R. aquatilis contains the pqqD and pqqE genes. Mutants of R. aquatilis HX2 showed that a lack of antibacterial activity was due to a Tn5 insertion in the pqqE gene, which prevented synthesis of the PQQ [9]. The aim of our study is to investigate the individual roles of pqqA and pqqB genes in R. aquatilis HX2 PQQ biosynthesis. The relevance of these genes to growth and beneficial activities of HX2 are assessed with respect to synthesis of OAs, antibacterial activity, mineral phosphate solubilization, biological control of crown gall disease, and plant growth promotion.

General Genetic Techniques
Isolation of genomic DNA from strain HX2 and plasmid DNA from E. coli were performed according to standard procedures [27]. Restriction enzyme digestions were performed as recommended by the suppliers (TaKaRa, Japan) and ligations were carried out using T4 DNA ligase (TaKaRa, Japan). Gel electrophoresis was performed in 0.8-1.0% agarose gels. For cloning purposes, Ex2 Taq DNA polymerase (TaKaRa, Japan) was used to PCR amplify inserts and Taq DNA polymerase (TaKaRa, Japan) was used for PCR amplification in test reactions (e.g., colony PCR). DNA sequencing was performed by Invitrogen Life Technologies (Beijing, China) and analyzed by using the National Center for Biotechnology Information BLAST server (http://www.ncbi.nlm.nih.gov/BLAST). The partial genome sequence of R. aquatilis HX2 (accession number CP003403-6) was used for primer design [28]. Primers used in this study are listed in Table 1.

Construction of pqqA and pqqB In-frame Deletion Mutants
In-frame nonpolar deletions of pqqA and pqqB were constructed utilizing a twostep homologous recombination strategy as described previously [29]. Primers were designed based on sequences upstream and downstream of either pqqA or pqqB and were used to PCR amplify fragments from the genome of HX2 (Table 1). Briefly, primers DpqqA-L1 and DpqqA-L2 with sites for Sal I and BamH I, were used to amplify a 1132 bp region upstream of the pqqA open reading frame (ORF). A 1553 bp fragment, created by primers DpqqA-R1 and DpqqA-R2 containing BamH I and EcoR I sites, is downstream of the pqqA ORF ( Table 1). The standard PCR reactions involved 15 min at 94˚C, then 35 cycles of 45 s at 94˚C, 40 s at 66˚C and 1 min at 72˚C, and final extension at 72˚C for 10 min. Similarly fragments flanking pqqB were amplified. These included a 982 bp upstream fragment that included the first 193 codons of the pqqB ORF, generated from primers DpqqB-L1 and DpqqB-L2 with sites EcoR I and BamH I and a 958 bp downstream fragment including the last 171 codons of the pqqB ORF from primers DpqqB-R1 and DpqqB-R2 with BamH I and Sal I sties. These PCR reactions were carried out at 94˚C for 15 min, followed by 35 cycles of 94˚C for 45 s, 71˚C for 2 min, and a final extension at 72˚C for 10 min. After being digested with appropriate restriction enzymes, the DpqqA-L/R and DpqqB-L/R fragments were ligated into pBSNot6 [30] creating pBSNot6DpqqA and pBSNot6DpqqB. Then, the approximately 2.7 kb Not I fragment from pBSNot6DpqqA and the approximately 2.0 kb Not I fragment from pBSNot6DpqqB, including a pqqA gene with 32 bp deletion and a pqqB gene with 564 bp deletion, were lifted and ligated into pSR47S [31] to obtain pSR47SDpqqA and pSR47SDpqqB. The two suicide plasmids were transformed into E. coli DH5a (l-pir) by heat shock and were mobilized from DH5a (l-pir) into the wild-type R. aquatilis strain HX2 by triparental mating with helper strain DH5a carrying plasmid pRK600 [32]. Exconjugants were selected on AB minimal agar plates containing kanamycin and second recombination events were selected according to the methods previous described [9]. The mutants of HX2DA and HX2DB were each screened from approximately 1500 first recombination clones for the absence of kanamycin resistance. A fragment 33 bp from the pqqA 72 bp gene was deleted in mutant HX2DA and 564 bp of the 912 bp pqqB gene was deleted in HX2DB. The disruption of each gene and the absence of the vector within the genome of these mutants were confirmed by PCR with primers DpqqA-L1/DpqqA-R2 and DpqqB-L1/DpqqB-R2 and sequencing analysis.

Genetic Complementation of the pqqA and pqqB Mutants
To complement the pqq mutants an 8.0 kb BamH I fragment containing the entire pqqABCDEF operon of HX2 was cloned into the broad host range vector pRK415G [33], resulting in the complementation plasmid pCH15 [9]. The plasmid pCH15 was mobilized into the HX2nA or HX2nB strain by triparental mating, and the complement strains CHX2nA and CHX2nB were created.

Antibiosis Test in Vitro and Biocontrol of Crown Gall Disease in Greenhouse
The antagonist HX2 and its derivative strains were tested for in vitro antibiosis against pathogenic strain A. vitis K308 via a modified Stonier's method described by Chen et al [3]. Biocontrol activity assays were performed on sunflower (Helianthus annuus L.) stems with two true leaves grown in a greenhouse according to a previously described method [3]. Briefly, a suspension of pathogenic agrobacterial strain K308 (ca. 2610 8 CFU ml 21 ) was mixed with an equal volume of HX2 or its derivative strains suspension (ca. 2610 8 CFU ml 21 ). A 10 ml drop of this mixture was injected into a 1.0 cm longitudinal incision in sunflower stem. The inoculation site was wrapped with Parafilm. Gall formation was observed, and the gall was excised and weighed 15 days after inoculation. Sterile buffered saline (SBS, 0.85% NaCl) was applied as a negative control and K308 mixed with SBS served as a positive control. The effectiveness index (EI) was calculated using the following formula: EI (%) 5 [(C-T)/C]6100, where C is the mean fresh weight of the crown gall tumor of the positive control group and T is the mean fresh weight of the crown gall tumor in the treated group. The assay was performed with 4 replicates and 10 plants were used per treatment.

Determination of Mineral Phosphate Solubilization and pH
HX2 and derivative strains were inoculated and shaken in PDB at 28˚C, 170 rpm for 48 h. Ten microliters of bacteria culture were dropped on a sterile filter paper (diameter, 5 mm), placed in the middle of agar plates containing the differential medium, national botanical research institute's phosphate (NBRIP). The NBRIP growth medium contains (per liter): glucose, 10 g; Ca 3 (PO 4 ) 2 , 5 g; MgCl 2 .6H 2 O, 5 g; MgSO 4 .7H 2 O, 0.25 g; KCl, 0.2 g; (NH 4 ) 2 SO 4 , 0.1g [34] with 18 g agar. The phosphate-solubilizing halo diameter was measured after the plates were incubated at 28˚C for 7 days.
HX2 and derivative strains were cultured in liquid NBRIP at 28˚C for 7 days to detect the soluble phosphorus (P) in the medium. Two-milliliters of bacterial supernatants were collected by centrifuge at 12,000 g for 5 min every other day. Soluble phosphate in the culture supernatants was detected using the Molybdenum-blue method [35]. Simultaneously, pH of the corresponding NBRIP medium was detected using a pH monitor (Mettler Toledo FE20, Shanghai, China). This assay was performed in triplicate with 3 replicates per treatment. Results from the third experiment are reported here.

Greenhouse Experiments for Plant Growth Promotion
Soil and coarse sand (0.35-0.5 mm) (1:1, w/w) were air-dried, passed through a sieve (2 mm), and sterilized by autoclaving at 121˚C for 2 h before filling the pots. The soil texture is sandy loam (70.8% sand, 26.9% silt, and 2.3% clay). The soil type is a calcaric cambisol according to the FAO/UNESCO soil map of the world. Maize caryopses (Zea may L., Zhengdan 958, Henan Academy of Agricultural Sciences, China) were surface sterilized with 70% ethanol for 30 sec, washed with sterilized distilled water for three times, and germinated in a sterilized Petri dish (270 mm) for 2 days. The bacteria strains were cultured with PDB at 28˚C for 2 days and diluted to OD 600 of 0.6 (ca. 2610 8 CFU ml 21 ). Maize seeds were coated with the diluted bacteria suspension or control (PDB) for 3 h before sowing. Coated maize caryopses were cultivated individually in pots (diameter 180 mm, length 160 mm) containing 750 g of soil (P 2 O 5 , 6.53 mg kg 21 ; soil organic matter, 18

Detection of PQQ
The presence of PQQ in culture supernatants was determined as described previously [9]. In brief, bacterial strains were grown at 28˚C in AB minimal medium for 48 h, cell cultures were then mixed with methanol at a 1:9 ratio (v/v). Precipitated material was removed by centrifugation (12,000 g, 15 min) and the methanol was evaporated with rotary evaporator. The sample was acidified with HCl to pH 2.0 and loaded onto a Sep-Pak C 18 cartridge (Agilent Technologies, USA). The cartridge was washed with 20 ml of 2 mM HCl and then PQQ was eluted with 70% methanol. To identify the PQQ peak, 200 ml of the sample was mixed with 100 ml of 0.2 M Na 2 B 4 O 7 buffer, adjusted to pH 8.0 with HCl, and mixed with 90 ml 0.5% acetone, then incubated for 30 min at room temperature. Reverse-phase high-performance liquid chromatography (RP-HPLC) was carried out using a Shimadzu LC-6A HPLC system with a fluorescence detector on ZORBAX SB-C 18 columns (4.66250 mm, 5 mm; Agilent Technologies, USA) eluted with 27% methanol and 0.4% H 3 PO 4 at a flow rate of 0.8 ml min 21 . The excitation and detection wavelengths of the fluorescence detector were set at 360 nm and 480 nm.

Identification of Organic Acid
Organic acids (OA) produced by bacterial strains were detected as described previously with some modifications [36]. HX2 and derivative strains were cultivated in NBRIP medium at 28˚C for 7 days. The culture was centrifuged at 12,000 g for 5 min and filtrated through a 0.22 mm filter (Pall Corporation, USA). Twenty microliters of filtrates were injected to high-performance liquid chromatography (HPLC) (Waters 2998, USA) equipped with ZORBAX SB-C 18 columns (4.66250 mm, 5 mm; Agilent Technologies, USA). The chromatogram class (NH 4 ) 2 HPO 4 (0.5%, w/v) with a pH of 2.81 was used as mobile phase at flow rate of 0.4 ml min 21 . UV absorption was routinely monitored at a wavelength of 214 nm. Five types of OA served as standards for all samples; gluconic acid (GA), lactic acid (LA), citric acid (CA), succinic acid (SA) and propionic acid (PA).

Statistical analysis
In this study, mean values among treatments were compared by Duncan's Multiple Range test at R,0.05. Analysis of variance (ANOVA) was performed on the data using SAS software (version 8.2; SAS, Inc., Cary, NC).

Antibiotic Production and Biocontrol are Associated with pqqA and pqqB
The pqqA and pqqB in-frame deletion mutants HX2DA and HX2DB lacked the ability to inhibit growth of A. vitis K308 on PDA (Table 2). Strains CHX2DA and CHX2DB, which were HX2DA and HX2DB strains complemented by plasmids containing the pqq gene cluster, displayed phenotypes of A. vitis K308 inhibition in vitro (Table 2). Strain HX2DA and HX2DB showed lesser biocontrol activity to crown gall disease on sunflower than wild-type strain HX2 and complement strain CHX2DA and CHX2DB (Table 2). Strain HX2DB showed higher biocontrol efficiency compared to strain HX2DA and EI of crown gall disease treated by HX2DB significantly increased by 41% compared to HX2DA treatment (P,0.05) ( Table 2 and Fig. 1). The abilities to produce antibacterial substance, inhibit growth of A. vitis K308 in vitro, and suppress gall development on sunflower were fully restored in the complemented strain CHX2DA and CHX2DB, respectively ( Table 2 and Fig. 1).

Genes pqqA and pqqB are Related to Phosphate Solubilization and Plant Growth Promotion
Based on the growth conditions of the bacteria strains in NBRIP with agar after 7day incubations, it was shown that there was a significant decrease (P,0.05) of phosphate-solubilizing halo diameters when either the pqqA or pqqB gene was disrupted (Table 3). When HX2DA or HX2DB was complemented with the pqq gene cluster (CHX2DA or CHX2DB), phosphate solubilization on NBRIP plates was restored. Wild-type HX2 excreted high levels of OA, especially GA and manifested much stronger MPS ability compared to strains HX2DB and HX2DA ( Table 2). Strains CHX2DA and CHX2DB restored intrinsic MPS ability and secreted similar levels of GA compared to the wild-type strain HX2. The phosphate-solubilizing halo diameters of R.aquatilis HX2, CHX2DA, and CHX2DB colonies were about two-fold greater than that of HX2DA and HX2DB strains ( Table 3). The phosphate-solubilizing halo of complemented strains CHX2DA and CHX2DB were, in-fact, even greater than those of the wild type strain HX2 (significant at P,0.05) ( Table 3).
To confirm that pqqA and pqqB were related to phosphate solubilization, additional tests were performed to quantify soluble P and lower culture pH in liquid media (Table 3 and S1 Figure). After 7-day incubations, the concentration of soluble P in culture solutions treated with CHX2DA, CHX2DB were 450.3 mg l 21 and 465.9 mg l 21 , which were 3.7 and 3.5 times greater than mutant HX2DA and HX2DB (Table 3). There was significant decrease (P,0.05) in soluble P concentration between strains with a disrupted pqqA or pqqB gene compared to the wild type strain HX2. At the same time, the pH value of culture solutions increased from 3.45 to 4.68 and 4.53 when the pqqA or pqqB gene was disrupted ( Table 3). Disruption of pqqA resulted in the lowest concentration of soluble P (99.7 mg l 21 ) of the derivative strains tested and also had the highest culture pH (4.68). The largest quantity of soluble P (465.9 mg l 21 ) was present in the   CHX2DB culture, corresponding to the lowest pH (3.38). There was a significant negative correlation (R 2 520.97, P,0.01) between pH and soluble P. The effects of strains HX2DA and HX2DB on plant growth promotion were monitored using pot experiments in the greenhouse. The details of the plant length, fresh and dry weight, total P and soluble P at 42 days post-inoculations are listed in Table 4. All plant growth parameters and conditions measured here were improved in the presence of HX2 as compared to the negative control. Specifically, maize plant length increased by 48%, fresh weight was 255% greater, dry weigh increased by 168%, plant total P was 67% greater, and there was 182% more soluble P in the soil. The effects of HX2 and derivative strains on plant growth promotion are significant, as compared to a blank control. However, HX2, CHX2DA, and CHX2DB treatments were more effective in enhancing maize length, fresh weight, dry weight, total P and soluble P in soil than HX2DA or HX2DB mutant treatments. There was no significant difference in length or fresh weight among HX2DA and HX2DB mutant treatments. However, there was a significant difference between HX2DB treatment and HX2DA treatments in which HX2DB displayed 19% greater dry plant weight, 17% more total P in maize, and 18% more soluble P in soil than HX2DA. In contrast to the negative control, the effect of HX2DB also shows significant improvement with respect to plant dry weight and soil soluble P where as HX2DA treatment only resulted in significantly greater soil soluble P. There was a significant positive correlation (R 2 50.87, P,0.01)between fresh weight and soil soluble P. Soil soluble P and NBRIP media Table 3. Diameter of phosphate-solubilizing zone, pH and soluble P observed after 7 d incubation in NBRIP.

Strains
Phosphate solubilizing halo diameter (mm) Soluble P (mg l 21 ) pH of medium Gluconic acid (g l 21 ) Role of pqqA and pqqB Genes in Rahnella aquatilis soluble P treated by the strains was positively correlated as well (R 2 50.72, P ,0.01).

Detection of PQQ
The level of PQQ production by HX2DA and HX2DB, determined with HPLC showed that pqqA and pqqB mutants lost the ability to biosynthesize PQQ ( Table 2 and Fig. 2). Non-detectable levels of PQQ were shown for each pqq mutant (HX2DA and HX2DB), but an average of 8.0 ng ml 21 and 7.9 ng ml 21 were detected in complemented strains CHX2DA and CHX2DB. However, PQQ produced by the complemented strains was significantly less than wild strain HX2 ( Table 2).

OA Production
Gluconic acid was the main OA produced by R. aquatilis HX2 which accounted for 94.3% of total OA production (Fig. 3). While HX2 produced GA and LA, HX2DA and HX2DB with their complemented strains only produced GA (Fig. 3). Strains HX2DA and HX2DB excreted low quantities of gluconic acid, as compared to all strains with intact pqq gene clusters (Table 3). HX2DB produced significantly more GA than did HX2DA (P,0.05). Furthermore, CHX2DB GA production was significantly greater than that of CHX2DA (P,0.05). Production of GA shows a significant negative correlation (r520.984, P,0.01) to pH of culture solution and is positively correlated (r50.973, P,0.01) to amount of soluble P in culture solutions.

Discussion
The most significant findings of this study are the different outcomes of pqqA and pqqB mutations on biocontrol capabilities, P solubilization, and plant-growth promotion by strain HX2. Although the pqqA gene product is redundant for the synthesis of PQQ in some bacteria, its availability was essential for PQQ biosynthesis in HX2 (Table 2) [37]. Of the derivative strains, HX2DA exhibited the lowest amounts of soluble P and excretion of organic acid and cultures had the Role of pqqA and pqqB Genes in Rahnella aquatilis highest pH. Notably, the HX2DB mutant displayed significantly greater GA production, soil soluble P and dry plant weight when compared to HX2DA (Table 3 and Table 4). This indicates a different role of pqqB in PQQ biosynthesis than that of pqqA. Furthermore, when HX2DA and HX2DB were complemented with the entire pqqABCDEF operon, the derivatives have additional pqq genes as compared to the wild-type strain. Forexample, CHX2DB now has one pqqB gene and duplicate pqqA genes. There are several differences between CHX2DA and CHX2DB. Specifically, CHX2DB produces significantly more GA than CHX2DA and the wild-type strain. However PQQ was not measured at higher levels for CHX2DB. All together, these findings indicate that pqqA is more important than pqqB in biocontrol capabilities, P solubilization, and plant-growth promotion by strain HX2. These results fall in line with those previously reported. Velterop et al. [22] found that the PqqA protein provided precursors for PQQ biosynthesis and that PqqB proteins facilitated the transport of PQQ across the cytoplasmic membrane into the periplasm in Klebsiella pneumonia. In this case PQQ biosynthesis was rate dependent on PqqA [22]. Hence, in HX2 the function of the pqqA g and pqqB genes may be similar to that in K. pneumonia. Many researchers have shown that the abilities of gram-negative bacteria, such as R. aquatilis, Pseudomonas cepacia, and Enterobacter intermedium, to solubilize insoluble phosphates were dependent upon pqq genes [38,39,40,41]. PQQ is Role of pqqA and pqqB Genes in Rahnella aquatilis necessary for the assembly of the GDH holoenzyme, which acts in the oxidation of glucose to GA [42,43]. The results described here indicate that the MPS ability of R. aquatilis HX2 was mainly determined by GA and LA production, and GA was significantly more important (Table 3 and Fig. 2). The correlation between quantities of GA and soluble P concentration indicates that GA production by HX2 effectively reduces the medium pH to increase MPS ability (Table 3, Fig 2  and S1 Figure). The main mechanism of insoluble phosphate dissolution in strain HX2 relates to glucose metabolism, in which GA is produced under the action of GDH utilizing PQQ as cofactor. Also, GA production will be greater if PQQ biosynthesis is increased. In this strain, pqqA and pqqB are both required for PQQ biosynthesis and to have wild-type levels of GA production. Although, it was at a greatly reduced level, strain HX2DA still produced GA but PQQ production was non-detectable. This brings to question if there is a GDH-PQQ independent pathway for GA catabolism in HX2.
Plant growth promotion was directly related to strain P solubilization ability in soil and culture medium ( Table 3 and Table 4). Maize growth promotion was greatest when treatments included HX2, CHX2DA, and CHX2DB, while HX2DA and HX2DB treatments demonstrated promotion to a significantly lesser extent. Likewise, strains HX2, CHX2DA and CHX2DB produced the highest amounts of GA, which released insoluble P into the soil solution (Table 4). This is consistent with previous findings where inoculation of phosphate solubilizing bacteria such as Serratia marcescens, Pseudomonas fluorescens and Bacillus spp. improved the phosphorous uptake of shoots and grains in maize and peanut plants [44,45,46]. Even with the greatly reduced GA production levels, HX2DA and HX2DB had benefits with respect to soil P-solubilization as compared to negative controls. It has been reported that plant growth promotion can be achieved by direct and indirect interaction between beneficial microbes and their host plants [47]. Given that HX2DA and HX2DB promoted plant growth, it is likely that strain HX2 has other plant-growth promoting properties in addition to phosphate solubilization.
Derivative strain HX2DB showed more effective biocontrol than did HX2DA ( Table 2). This is likely related to greater GA production by HX2DB. Gluconic acid can serve as an antifungal agent and has been associated with the regulation of other antimicrobial compounds, 2,4-diacethylphloroglucinol and pyoluteorin [11,48]. Without PQQ production, HX2DA and HX2DB mutants had greatly reduced abilities to produce GA and disrupted the ability to produce antibacterial substance, but they were still able to suppress tumor formation on sunflowers significantly (Fig. 1). Hence, it is likely that production antibacterial substance and GA is not the sole mechanism involved in HX2 biocontrol of crown gall disease.
Further work should be done to identify the potential additional plant growth promoting factors and antibacterial substances produced by HX2. More information will provide insight to optimized growing conditions or modifications to the strain which can lend it improved benefits to plants.